1. CPSC 231: Functions: Decomposition And Code Reuse This section of notes shows you how to write functions that can be used to: decompose large problems, and to reduce program size by creating reusable sections. slide 1

2. James Tam Example Programs Location (via the WWW): – http://pages.cpsc.ucalgary.ca/~tamj/231/examples/decomposition Location (via the CPSC UNIX network): – /home/231/examples/decomposition

3. James Tam Tip For Success: Reminder Look through the examples and notes before class. This is especially important for this section because the execution of these programs will not be sequential order. Instead execution will appear to ‘jump around’ so it will be harder to follow the examples if you don’t do a little preparatory work. Also it would be helpful to take notes that include greater detail: – For example: Literally just sketching out the diagrams that I draw without the extra accompanying verbal description that I provide in class probably won’t be useful to study from later.

4. James Tam Solving Larger Problems Sometimes you will have to write a program for a large and/or complex problem. One technique employed in this type of situation is the top down approach to design. – The main advantage is that it reduces the complexity of the problem because you only have to work on it a portion at a time.

5. Top Down Design 1.Start by outlining the major parts (structure) 2.Then implement the solution for each part My autobiography Chapter 1: The humble beginnings Chapter 2: My rise to greatness … Chapter 7: The end of an era Chapter 1: The humble beginnings It all started ten and one score years ago with a log-shaped computer work station… Image copyright unknown

6. James Tam Procedural Programming Applying the top down approach to programming. Rather than writing a program in one large collection of instructions the program is broken down into parts. Each of these parts are implemented in the form of procedures (also called “functions”, “procedures” or “methods” depending upon the programming language).

7. Procedural Programming Main tasks to be fulfilled by the program Important subtask #1 Important subtask #2 Important subtask #3 Function #1 …Etc. Function #2 Function #3 …Etc. When do you stop decomposing and start writing functions? No clear cut off but use the “Good style” principles (later in these notes) as a guide e.g., a function should have one well defined task and not exceed a screen in length.

8. Decomposing A Problem Into Functions Break down the program by what it does (described with actions/verbs or action phrases). Eventually the different parts of the program will be implemented as functions.

9. Example Problem Design a program that will perform a simple interest calculation. The program should prompt the user for the appropriate values, perform the calculation and display the values onscreen.

10. Example Problem Design a program that will perform a simple interest calculation. The program should prompt the user for the appropriate values, perform the calculation and display the values onscreen. Action/verb list: – Prompt – Calculate – Display

11. Top Down Approach: Breaking A Programming Problem Down Into Parts (Functions) Calculate Interest Get informationDo calculationsDisplay results

12. Things Needed In Order To Use Functions Function definition – Instructions that indicate what the function will do when it runs. Function call – Actually running (executing) the function. – You have already done this second part many times because up to this point you have been using functions that have already been defined by someone else e.g., print(), input()

13. Functions (Basic Case: No Arguments) Function call Function definition

14. Defining A Function Format: def (): body 1 Example: def displayInstructions(): print ("Displaying instructions on how to use the program") 1 Body = the instruction or group of instructions that execute when the function executes (when called). The rule in Python for specifying the body is to use indentation.

15. Calling A Function Format: () Example: displayInstructions()

16. James Tam Quick Recap: Starting Execution Point The program starts at the first executable instruction that is not indented. In the case of your programs thus far all statement have been un-indented (save loops/branches) so it’s just the first statement that is the starting execution point. But note that the body of functions MUST be indented in Python. HUMAN_CAT_AGE_RATIO = 7 age = input("What is your age in years: ") catAge = age * HUMAN_CAT_AGE_RATIO …

17. James Tam Functions: An Example That Puts Together All The Parts Of The Easiest Case Name of the example program: firstExampleFunction.py def displayInstructions(): print("Displaying instructions") # Main body of code (starting execution point, not indented) displayInstructions() print("End of program")

18. James Tam Name of the example program: firstExampleFunction.py def displayInstructions(): print("Displaying instructions") # Main body of code (starting execution point) displayInstructions() print("End of program") Functions: An Example That Puts Together All The Parts Of The Easiest Case Function definition Function call

19. James Tam Defining The Main Body Of Code As A Function Rather than defining instructions outside of a function the main starting execution point can also be defined explicitly as a function. (The previous program rewritten to include an explicit start function) “ firstExampleFunction2.py ” def displayInstructions(): print ("Displaying instructions") def start(): displayInstructions() print("End of program ") Important: If you explicitly define the starting function then do not forgot to explicitly call it! start () Don’t forget to start your program! Program starts at the first executable un-indented instruction

20. James Tam Stylistic Note By convention the starting function is frequently named ‘ main() ’ or in my case ‘start()’. def main(): OR def start(): This is done so the reader can quickly find the beginning execution point.

21. James Tam New Terminology Local variables: are created within the body of a function (indented) Global constants: created outside the body of a function. (The significance of global vs. local is coming up shortly). HUMAN_CAT_AGE_RATIO = 7 def getInformation(): age = input("What is your age in years: ") catAge = age * HUMAN_CAT_AGE_RATIO Global constant Local variables

22. James Tam Creating Your Variables Before: all statements (including the creation of a variables) occur outside of a function Now that you have learned how to define functions, ALL your variables must be created with the body of a function. Constants can still be created outside of a function (more on this later). HUMAN_CAT_AGE_RATIO = 7 age = input("What is your age in years: ") catAge = age * HUMAN_CAT_AGE_RATIO … HUMAN_CAT_AGE_RATIO = 7 def getInformation(): age = input("What is your age in years: ") catAge = age * HUMAN_CAT_AGE_RATIO ‘Outside’: OK for constants only Inside function body: all variables must be here

23. Variables are memory locations that are used for the temporary storage of information. num = 888 Each variable uses up a portion of memory, if the program is large then many variables may have to be declared (a lot of memory may have to be allocated to store the contents of variables). What You Know: Declaring Variables 888 num RAM

24. What You Will Learn: What Is The Significance Of Being ‘Local’ To minimize the amount of memory that is used to store the contents of variables only create variables when they are needed (“allocated”). When the memory for a variable is no longer needed it can be ‘freed up’ and reused (“de-allocated”). To design a program so that memory for variables is only allocated (reserved in memory) as needed and de-allocated when they are not (the memory is free up) variables should be declared as local to a function. (There’s an even better reason for making variables local coming up later ‘side effects’)

25. James Tam What You Will Learn: How To Work With Locals Function call (local variables get allocated in memory) The program code in the function executes (the variables are used to store information needed for the function) Function ends (local variables get de-allocated in memory)

26. Reminder: Where To Create Local Variables def (): Example: def fun(): num1 = 1 num2 = 2 Somewhere within the body of the function (indented part)

27. James Tam Working With Local Variables: Putting It All Together Name of the example program: secondExampleFunction.py def fun(): num1 = 1 num2 = 2 print(num1, " ", num2) # start function fun() Variables that are local to function ‘ fun ’

28. James Tam Another Reason For Creating Local Variables To minimize side effects (unexpected changes that have occurred to variables after a function has ended e.g., a variable storing the age of the user accidentally takes on a negative value). To visualize the potential problem: imagine if all variables could be accessed anywhere in the program (not local). Memory Fun1 () Fun2 () Fun3 () variable ???

29. James Tam Recall: local variables only exist for the duration of a function. After a function ends the local variables are no longer accessible. Benefit: reduces accidental changes to local variables. Local Variables def fun(): x = 7 y = 13 def start(): a = 1 fun() # x,y # inaccessible RAM 7 x 13 y Memory: ‘ fun ’ 1 a Memory: ‘ start ’

30. James Tam Recall: local variables only exist for the duration of a function. After a function ends the local variables are no longer accessible. Benefit: reduces accidental changes to local variables. Local Variables def fun(): x = 7 y = 13 RAM def start(): a = 1 fun() # x,y # inaccessible 7 x 13 y Memory: ‘ fun ’ 1 a Memory: ‘ start ’ Not possible (good!)

31. James Tam New Problem: Local Variables Only Exist Inside A Function def display (): print ("") print ("Celsius value: ", celsius) print ("Fahrenheit value :", fahrenheit) def convert (): celsius = float(input ("Type in the celsius temperature: ")) fahrenheit = celsius * 9 / 5 + 32 display () What is ‘ celsius ’??? What is ‘ fahrenheit ’??? New problem: How to access local variables outside of a function? Variables celsius and fahrenheit are local to function ‘ convert() ’

32. James Tam One Solution: Parameter Passing Passes a copy of the contents of a variable as the function is called: convert celsius fahrenheit Parameter passing: communicating information about local variables (via arguments) into a function display Celsius ? I know that value! Fahrenheit ? I know that value!

33. Parameter Passing: Past Usage You did it this way so the function ‘knew’ what to display: age = 27 # Pass copy of 27 to # print() function print(age) You wouldn’t do it this way: age = 27 # Nothing passed to print # Function print() has # no access to contents # of ‘age’ print() # Q: Why doesn’t it # print my age?! # A: Because you didn’t # tell it to!

34. James Tam Parameter Passing (Function Definition) Format: def (,..., ): Example: def display(celsius, fahrenheit):

35. James Tam Parameter Passing (Function Call) Format: (,..., ) Example: display(celsius, fahrenheit)

36. James Tam Memory And Parameter Passing Parameters passed as arguments into functions become variables in the local memory of that function. def fun(num1): print(num1) num2 = 20 print(num2) def start(): num1 = 1 fun(num1) start() num1 : local to start Parameter num1 : local to fun num2 : local to fun

37. James Tam Useful for visualizing the layout of function calls in a large and complex program. Format: Example: def start(): age = float(input()) print(age) Structure Charts Function being called Calling function Function being called input start float print age

38. James Tam Structure Chart: temperature.py To reduce clutter most structure charts only show functions that were directly implemented by the programmer. introduction start convert display celsius fahrenheit Args(celsius,fahrenheit)

39. James Tam Parameter Passing: Putting It All Together Name of the example program: temperature.py def introduction (): print (""" Celsius to Fahrenheit converter ------------------------------- This program will convert a given Celsius temperature to an equivalent Fahrenheit value. """)

40. James Tam Parameter Passing: Putting It All Together (2) def display (celsius, fahrenheit): print ("") print ("Celsius value: ", celsius) print ("Fahrenheit value:", fahrenheit) def convert (): celsius = float(input ("Type in the celsius temperature: ")) fahrenheit = celsius * 9 / 5 + 32 display (celsius, fahrenheit) # start function def start (): introduction () convert () start ()

41. James Tam A parameter is copied into a local memory space. Parameter Passing: Important Recap! # Inside function convert() display(celsius, fahrenheit) # Function call # Inside function display def display(celsius, fahrenheit): # Function # definition Make copy Data copy Separate RAM -34 celsius -29.2 fahrenheit Memory: ‘ convert ’ -34 celsius -29.2 fahrenheit Memory: ‘ display ’ Separate

42. James Tam The Type And Number Of Parameters Must Match! Correct : def fun1(num1, num2): print(num1, num2) def fun2(num1, str1): print(num1, str1) # start def start(): num1 = 1 num2 = 2 str1 = "hello" fun1(num1, num2) fun2(num1, str1) start() Two numeric parameters are passed into the call for ‘ fun1() ’ which matches the two parameters listed in the definition for function ‘ fun1() ’ Two parameters (a number and a string) are passed into the call for ‘fun2()’ which matches the type for the two parameters listed in the definition for function ‘fun2()’

43. James Tam A Common Mistake: The Parameters Don’t Match Incorrect  : def fun1(num1): print (num1, num2) def fun2(num1, num2): num1 = num2 + 1 print(num1, num2) # start def start(): num1 = 1 num2 = 2 str1 = "hello" fun1(num1, num2) fun2(num1, str1) start() Two numeric parameters are passed into the call for ‘ fun1() ’ but only one parameter is listed in the definition for function ‘ fun1() ’ Two parameters (a number and a string) are passed into the call for ‘ fun2() ’ but in the definition of the function it’s expected that both parameters are numeric.

44. James Tam Documenting Functions Python doesn’t require the type to be specified in the parameter list. Therefore the number and type of parameters should be specified in the documentation for the function. # display(float,float) def display(celsius, fahrenheit):

45. James Tam Yet Another Common Mistake: Not Declaring Parameters You wouldn’t do it this way with pre-created functions: def start(): print(num) So why do it this way with functions that you define yourself: Etc. (Assume fun() has been defined elsewhere in the program) # start def start(): fun(num) start() What is ‘ num ’? It has not been declared in function ‘ start() ’ What is ‘ num ’? It has not been created in function ‘ start() ’ # start (corrected) def start(): num = fun(num) start()

46. Scope The scope of an identifier (variable, constant) is where it may be accessed and used. In Python 1 : – An identifier comes into scope (becomes visible to the program and can be used) after it has been declared. – An identifier goes out of scope (no longer visible so it can no longer be used) at the end of the indented block where the identifier has been declared. 1 The concept of scoping (limited visibility) applies to all programming languages. The rules for determining when identifiers come into and go out of scope will vary with a particular language.

47. Scope: An Example def fun1(): num = 10 # statement # End of fun1 def fun2(): print(num) : : ‘ num ’ comes into scope (is visible and can be used) (End of function): ‘ num’ goes out of scope, no longer accessible Scope of num Num is no longer in scope Error: ‘ num’ is an unknown identifier

48. Scope: A Variant Example def fun1(): num = 10 # statement # End of fun1 def fun2(): fun1() num = 20 : : What happens at this point? Why?

49. New Problem: Results That Are Derived In One Function Only Exist While The Function Runs def calculateInterest(principle, rate, time): interest = principle * rate * time # start principle = 100 rate = 0.1 time = 5 calculateInterest (principle, rate, time) print(“Interest earned $”, interest) Stored locally interest = 50 Problem: Value stored in interest cannot be accessed here

50. Solution: Have The Function Return Values Back To The Caller def calculateInterest(principle, rate, time): interest = principle * rate * time return(interest) # start principle = 100 rate = 0.1 time = 5 interest = calculateInterest(principle, rate, time) print (“Interest earned $”, interest) Variable ‘ interest ’ is still local to the function. The value stored in the variable ‘ interest ’ local to ‘ calculateInterest() ’ is passed back and stored in a variable that is local to the “start function”.

51. James Tam Remember that local variables only exist for the duration of a function. Function Return Values (1) def calculateArea(): w = int(input()) l = int(input()) a = w * l def main(): calculateArea() print(area ) RAM Memory: ‘ main ’ w l Memory: ‘ calculateArea ’ a

52. James Tam w l Memory: ‘ calculateArea ’ a After a function has ended local variables are ‘gone’. Function Return Values (2) def calculateArea(): w = int(input()) l = int(input()) a = w * l RAM Memory: ‘ main ’ Area? (‘a’ no longer exists) def main(): calculateArea() print(area )

53. James Tam w l Memory: ‘ calculateArea ’ a Function Return Values (3) Function return values communicate a copy of information out of a function (back to the caller) just as the function ends. def calculateArea(): w = int(input()) l = int(input()) a = w * l RAM Memory: ‘ main ’ return(a) area The return statement passes back a copy of the value stored in ‘ a ’ Copy of a’s data def main(): area = calculateArea() print(area )

54. Using Return Values Format (Single value returned) 1 : return( ) # Function definition = ()# Function call Example (Single value returned) 1 : return(interest)# Function definition interest = calculateInterest # Function call (principle, rate, time) 1 Although bracketing the return value isn’t required when only a single value is returned it’s still recommended that you get in the habit of doing it because it is required for ‘multiple’ return values. The actual details about the difference between returning a single vs. ‘multiple’ values will be covered in the ‘composites’ section.

55. Using Return Values Format (Multiple values returned): # Function definition return(,...) # Function call,... = () Example (Multiple values returned): # Function definition return(principle, rate, time) # Function call principle, rate, time = getInputs(principle, rate, time)

56. James Tam Structure Chart: interest.py introduction start getInputs calculate display principle rate time interest amount principle rate time interest amount Return (principle,rate,time) Args (principle,rate,time) Return (interest,amount) Args (principle,rate,time, interest,amount)

57. Using Return Values: Putting It All Together Name of the example program: interest.py def introduction(): print(""" Simple interest calculator ------------------------------- With given values for the principle, rate and time period this program will calculate the interest accrued as well as the new amount (principle plus interest). """)

58. Using Return Values: Putting It All Together (2) def getInputs(): principle = float(input("Enter the original principle: ")) rate = float(input("Enter the yearly interest rate %")) rate = rate / 100 time = input("Enter the number of years that money will be invested: ") time = float(time) return(principle, rate, time) def calculate(principle, rate, time): interest = principle * rate * time amount = principle + interest return(interest, amount)

59. Using Return Values: Putting It All Together (3) def display(principle, rate, time, interest, amount): temp = rate * 100 print("") print("Investing $%.2f" %principle, "at a rate of %.2f" %temp, "%") print("Over a period of %.0f" %time, "years...") print("Interest accrued $", interest) print("Amount in your account $", amount)

60. Using Return Values: Putting It All Together (4) # start function def start(): principle = 0 rate = 0 time = 0 interest = 0 amount = 0 introduction () principle, rate, time = getInputs() interest, amount = calculate(principle, rate, time) display(principle, rate, time, interest, amount) start()

61. James Tam Stylistic Note Creating variables all at once at the start of a function. def start(): principle = 0 rate = 0 time = 0 interest = 0 amount = 0 introduction () principle, rate, time = getInputs() interest, amount = calculate(principle, rate, time) display(principle, rate, time, interest, amount) start() Not syntactically required but a stylistic approach

62. James Tam Return And The End Of A Function A function will immediately end and return back to the caller if: 1.A return statement is encountered (return can be empty “ None ”) def convert(catAge): if (catAge < 0): print(“Can’t convert negative age to human years”) return() # Explicit return to caller (return # statement) else: :: 2.There are no more statements in the function. def introduction(): print() print("TAMCO INC. Investment simulation program") print("All rights reserved") print() # Implicit return to caller (last statement)

63. James Tam Documenting Functions Similar to specifying the function arguments, the type of the return values should also be documented. Example: # calculate # returns(float,float) def calculate(principle, rate, time):

64. James Tam Another Common Mistake: Not Saving Return Values (Pre-Created Functions) You would typically never use the input() function this way (Function return value not stored) input(“Enter your name”) print(name) (Function return value should be stored) name = input(“Enter your name”) print(name)

65. Yet Another Common Mistake: Not Saving Return Values (Your Functions) Just because a function returns a value does not automatically mean the return value will be usable by the caller of that function. def fun(): return(1) Function return values must be explicitly saved by the caller of the function. def calculateArea(length,width): area = length * width return(area) # Start: error area = 0 calculateArea(4,3) print(area) This value has to be stored or used in some expression by the caller # Start: fixed area = 0 area = calculateArea (4,3) print(area)

66. James Tam Parameter Passing Vs. Return Values Parameter passing is used to pass information INTO a function. – Parameters are copied into variables that are local to the function. def start(): num = int(input("Enter number: ")) absolute(num) Memory: start num -10 Memory: absolute num -10 def absolute(num): etc.

67. James Tam Parameter Passing Vs. Return Values Return values are used to communicate information OUT OF a function. – The return value must be stored in the caller of the function. Memory: start num 3 Memory: square result 9 def square(num): result = num * num return(result) num 3 Parameter result 9 Return value def start(): num = int(input("Enter number: ")) result = square(num) print(result)

68. Global Scope Identifiers (constants or variables) that are declared within the body of a function have a local scope (the function). def fun (): num = 12 # End of function fun Identifiers (constants or variables) that are created outside the body of a function have a global scope (the program). num = 12 def fun1 (): # Instructions def fun2 (): # Instructions # End of program Scope of num is the function Scope of num is the entire program

69. Global Scope: An Example Name of the example program: globalExample1.py num1 = 10 def fun(): print(num1) def start(): fun() print(num2) num2 = 20 start()

70. Global Variables: General Characteristics You can access the contents of global variables anywhere in the program. In most programming languages you can also modify global variables anywhere as well. – This is why the usage of global variables is regarded as bad programming style, they can be accidentally modified anywhere in the program. – Changes in one part of the program can introduce unexpected side effects in another part of the program. – So unless you have a compelling reason you should NOT be using global variables but instead you should pass values as parameters. Unless you are told otherwise using global variables can affect the style component of your assignment grade. Global constants are acceptable and are commonly used.

71. Global Variables: Python Specific Characteristic Name of the example program: globalExample2.py num = 1 def fun(): num = 2 print(num) def start(): print(num) fun() print(num) start() Global Local created and displayed

72. James Tam Scoping Rules: Globals When an identifier is referenced (variable or constant) then: 1.First look in the local scope for the creation of the identifier: if found here then stop looking and use this identifier 2.If nothing exists at the local level then look globally def aFunction(): print(num) Reference to an identifier 2.Check globally Num = here? 1.Check locally Num = here?

73. Python Globals: ‘Read’ But Not ‘Write’ Access By default global variables can be accessed globally (read access). Attempting to change the value of global variable will only create a new local variable by the same name (no write access to the global, only the local is changed). num = 1 def fun(): num = 2 print(num) Prefacing the name of a variable with the keyword ‘ global ’ in a function will indicate references in that function will refer to the global variable rather than creating a local one. global Global num Local num

74. Globals: Another Example (‘Write’ Access Via The “ Global ” Keyword) Name of the example program: globalExample3.py num = 1 def fun(): global num num = 2 print(num) def start(): print(num) fun() print(num) start() Global References to the name ‘ num ’ now affect the global variable, local variable not created Global still changed after ‘ fun() ’ is done Global changed

75. James Tam What Level To Declare Variables Declare your variables as local to a function. When there are multiple levels of functions (a level is formed when one function calls another) then: – A variable should be created at the lowest level possible fun1 fun2Fun3(x,y) Need x,y here x,y Get and return x,y fun3 fun1 fun2 y,z x

76. James Tam Documenting Functions (As previously mentioned the documentation should include) – The type and number of arguments e.g., # fun(int,string) – The type and number of return values e.g., # returns(float,float,int) Additional documentation – Functions are a ‘mini’ program. – Consequently the manner in which an entire program is documented should also repeated in a similar process for each function: Features list. Limitations, assumptions e.g., if a function will divide two parameters then the documentation should indicate that the function requires that the denominator is not zero. (Authorship and version number may or may not be necessary for the purposes of this class although they are often included in actual practice).

77. James Tam Doc Strings A special form of documentation: – Characteristic 1: It allows for documentation to span multiple lines – Example: """ (triple double quotes) function: getInputs @getInputs(none) @returns(float,float,int) @Prompt the user for the inputs to the operation: principle, rate, time """ def getInputs():... return(principle, rate, time)

78. James Tam Doc Strings (2) – Characteristic 2: it can provide help as the program is running in Python’s interactive mode. – Example: program is stored in file called “ doc_strings.py ” – Interactive mode is invoked by typing “ python ” at the command line (no program name) """ function: getInputs @getInputs(none) @returns(float,float,int) @Prompt the user for the inputs to the operation: principle, rate, time """ def getInputs():... return(principle, rate, time) doc_strings.py Start interactive mode Viewing help (doc string)

79. James Tam Boolean Functions Return a Boolean value (true/false): “Asks a question” Typically the Boolean function will ‘ask the question’ about a parameter(s) Example: – Is it true that the string can be converted to a number? aString = input("Enter age: ") ageOK = isNum(aString) if (ageOK != True): print("Age must be a numeric value") else: # OK to convert the string to a number age = int(aString) # Boolean function def isNum(aString): # Returns (True # or False)

80. James Tam Good Style: Functions 1.Each function should have one well defined task. If it doesn’t then this may be a sign that the function should be decomposed into multiple sub-functions. a)Clear function: A function that squares a number. b)Ambiguous function: A function that calculates the square and the cube of a number. – Writing a function that is too specific makes it less useful (in this case what if we wanted to perform one operation but not the other). Also functions that perform multiple tasks can be harder to test.

81. James Tam Good Style: Functions (2) 2.(Related to the previous point). Functions should have a self descriptive action-oriented name (verb/action phrase or take the form of a question – the latter for functions that check if something is true): the name of the function should provide a clear indication to the reader what task is performed by the function. a)Good: drawShape(), toUpper() isNum(), isUpper() # Boolean functions: ask questions a)Bad: doIt(), go(), a()

82. James Tam Good Style: Functions (2) 3.Try to avoid writing functions that are longer than one screen in length. a)Tracing functions that span multiple screens is more difficult. 4.The conventions for naming variables should also be applied in the naming of functions. a)Lower case characters only. b)With functions that are named using multiple words capitalize the first letter of each word except the first (so called “camel case”) - most common approach or use the underscore (less common). Example: toUpper()

83. Functions Should Be Defined Before They Can Be Called! Correct def fun(): print("Works") # start fun() Incorrect  # Start fun() def fun(): print("Doesn't work") Function definition Function call Function definition Function call

84. James Tam Another Common Mistake Forgetting the brackets during the function call: def fun(): print("In fun") # start function print("In start") fun

85. James Tam Another Common Mistake Forgetting the brackets during the function call: def fun(): print("In fun") # start function print("In start") fun () The missing set of brackets do not produce a syntax/translation error

86. James Tam Another Common Problem: Indentation Recall: In Python indentation indicates that statements are part of the body of a function. (In other programming languages the indentation is not a mandatory part of the language but indenting is considered good style because it makes the program easier to read). Forgetting to indent: def start (): print ("start " ) start ()

87. James Tam Another Common Problem: Indentation (2) Inconsistent indentation: def start(): print("first") # Error: Unless this is the body of branch or loop print("second") start()

88. James Tam Yet Another Problem: Creating ‘Empty’ Functions def start(): start() Problem: This statement appears to be a part of the body of the function but it is not indented???!!!

89. James Tam Yet Another Problem: Creating ‘Empty’ Functions (2) def fun(): print() # start fun() A function must have at least one statement Alternative (writing an empty function: literally does nothing) def fun(): pass # start fun()

90. James Tam Testing Functions The correctness of a function should be verified. (“Does it do what it is supposed to do?”) Typically this is done by calling the function, passing in predetermined parameters and checking the result. Example: absolute_test.py def absolute(number): if (number < 0): result = number * -1 else: result = number return(result) # Test cases print(absolute(-13)) print(absolute(7)) Expected results: 13 7

91. James Tam Creating A Large Document Recall: When creating a large document you should plan out the parts before doing any actual writing. Chapter 1 Introduction Section 1.1 Section 1.2 Section 1.3 Conclusion Chapter 2 Introduction Section 2.1 Section 2.2 Section 2.3 Section 2.4 Conclusion Chapter 3 Introduction Section 3.1 Section 3.2 Conclusion Step 1: Outline all the parts (no writing) Section 1.1 It all started seven and two score years ago… Step 2: After all parts outlined, now commence writing one part at a time

92. James Tam Creating A Large Program When writing a large program you should plan out the parts before doing any actual writing. Step 1: Calculate interest (write empty ‘skeleton’ functions) def getInformation(): pass def doCalculations(): pass def displayResults(): pass Step 2: All functions outlined, write function bodies one-at- a-time (test before writing next function) def getInformation(): principle = int(input()) interest = int(input()) time = int(input()) return(principle,interest,time) # Simple test: check inputs # properly read and # returned p,r,t = getInformation() print(p,r,t)

93. Why Employ Problem Decomposition And Modular Design (1) Drawback – Complexity – understanding and setting up inter-function communication may appear daunting at first. – Tracing the program may appear harder as execution appears to “jump” around between functions. – These are ‘one time’ costs: once you learn the basic principles of functions with one language then most languages will be similar.

94. James Tam Why Employ Problem Decomposition And Modular Design (2) Benefit – Solution is easier to visualize and create (decompose the problem so only one part of a time must be dealt with). – Easier to test the program: Test one feature/function at a time (Testing multiple features increases complexity) – Easier to maintain (if functions are independent changes in one function can have a minimal impact on other functions, if the code for a function is used multiple times then updates only have to be made once). – Less redundancy, smaller program size (especially if the function is used many times throughout the program). – Smaller programs size: if the function is called many times rather than repeating the same code, the function need only be defined once and then can be called many times.

95. James Tam After This Section You Should Now Know How and why the top down approach can be used to decompose problems – What is procedural programming How to write the definition for a function How to write a function call How and why to declare variables locally How to pass information to functions via parameters How and why to return values from a function What is a Boolean function What is the difference between a local and a global variable. How to document a function

96. James Tam Copyright Notification “Unless otherwise indicated, all images in this presentation are used with permission from Microsoft.” slide 96